How can Total Productive Maintenance Enable Supply Network Resilience? A Case of Supply Networks in Utilities

(1)

Master Thesis Supply Chain Management

University of Groningen

How can Total Productive Maintenance Enable

Supply Network Resilience? A Case of Supply

Networks in Utilities

Benjamin G. H. Ward - S2525070

benjaminghward@me.com

Supervisor, University of Groningen: Dr. Kirstin Scholten Co-Assessor, University of Groningen: Prof. Dr. Dirk-Pieter van Donk

June 22nd, 2015 Word count: 11,707

Acknowledgments:

(2)

Abstract:

(5)

1. Introduction

At a basic level, lean and resilience appear to be conflicting ideas in a supply chain context owing to lean advocating the reduction of waste (Holweg, 2007; Shah & Ward, 2007), and resilience often being enabled by the addition of redundancies (Sheffi & Rice, 2005). Yet, as resilience is also the ability to grow, change and adapt when confronted with change (Fiksel, 2007), or more specifically, having the capacity to prepare for, respond to and recover from supply chain disturbances, it follows that lean supply chains also require a level of resilience, as no supply chain is immune to disturbances. This is especially pertinent as many authors have acknowledged that a focus on applying lean philosophies to supply chains has increased the risk of disturbances (Blackhurst, Dunn, & Craighead, 2011; Christopher & Peck, 2004; Christopher & Rutherford, 2004; Pettit, Croxton, & Fiksel, 2013; Ponomarov & Holcomb, 2009). Such a focus is indicative of a link between resilience and lean; however the relationship has largely been overlooked in academic research. It is therefore of interest to attempt to understand ways in which these seemingly conflicting goals interact in order to determine how lean philosophies can contribute to, or enable, resilience, rather than diminish it.

This study aims to investigate the relationship between Total Productive Maintenance (TPM) and supply chain resilience in particular. The antecedents of resilience have been discussed in literature (Christopher & Peck, 2004; Durach, Wieland, & Machuca, 2015; Ponomarov & Holcomb, 2009; Wieland & Wallenburg, 2013), often with relations to strategy (Wieland & Wallenburg, 2012), or balancing up- and down-stream influences on a supply chain (Christopher & Peck, 2004). TPM is a facet of lean philosophy and an approach to managing maintenance to ensure reliability in a system (Ahuja & Khamba, 2008). This research aims to investigate the potential beneficial effects on the maintenance management processes in an organisation leading to higher supply network resilience that can accrue through the application of the key tenets of TPM to the supply chain network. This relationship will be studied by way of a case study focussing on the supply network that relies heavily on maintenance to perform at a high service level.

(6)

on data from case study research the aim of this research is to answer the following question:

How can applying the use of Total Productive Maintenance build supply network resilience?

In answering the research question, this study makes three key contributions. Firstly, this research deconstructs TPM into key tenets to propose potential interaction with antecedents of resilience, and by extension, their contribution to enabling resilience. Secondly, this research explores these linkages in a supply network setting in order to understand whether TPM can enable resilience in such a setting. Lastly, this research makes a practical contribution in aiding utility-based organisations understand the impact TPM can have on the resilience of their maintenance and supply networks through the application of a case and the development of a decision-making artefact. A case company in this sector will be examined and a redesign of its supply network will be considered.

Following this introduction, a literature review will outline the current state of research on these topics and explain their potential interaction. This potential interaction will be clarified through the construction of a conceptual model. Subsequently a design methods methodology is applied to a single case study. Findings will discuss the outcomes of the case study with regards to the conceptual model and a discussion will elaborate the representativeness and applicability of the model and provide a base for future research. A final section will conclude.

2. Resilience and Total Productive Maintenance

2.1 Resilience

(7)

Robustness is seen as a proactive way to gain supply chain resilience (Durach et al., 2015), which aligns with the findings of Wieland & Wallenburg (2013), who argue that resilience is composed of two antecedents; robustness and agility. This line of thinking identifies two phases of resilience. The proactive phase, in which robustness is found, and the reactive phase, to which agility is crucial and which robustness contributes to by enhancing visibility and velocity through knock-on effects of a prepared supply chain. This view is further stated by Christopher & Peck (2004) whose definition of resilience implies flexibility, but also adaptability, should the desired post-disruptive state be different to the original state. By and large, research on resilience agrees that resilience itself is exemplified by the proactive and reactive phase (Christopher & Peck, 2004; Jüttner & Maklan, 2011; Pettit, Fiksel, & Croxton, 2010; Sheffi & Rice, 2005; Wieland & Wallenburg, 2013), despite various wording by different researchers. The Wieland & Wallenburg (2013) model of resilience is the primary model of resilience used in the research due to the clear-cut distinction between the proactive and reactive factors of resilience and its relative simplicity, allowing a deeper exploration of factors that have the potential to relate between resilience and TPM, unrestricted by the confines of more specific conceptualisations. 2.1.1 Robustness

(8)

to be quantified despite being successful. Similar findings are echoed by Brandon-Jones et al. (2014). To this end, Durach et al. (2015), provide a comprehensive review of intra- and inter-organisational antecedents to robustness, however it lacks applicability to services and the constructs are broad, conceptual and do not refer to consistent levels in the organisation.

As robustness has largely been identified as the proactive construct of resilience, it is important to conceptualise robustness by way of factors that emphasise this point. Wieland & Wallenburg (2013), identify anticipation and preparedness as the underlying mechanisms of robustness.

Anticipation consists of all actions taken to predict potential future events (Pettit et al., 2010), thus identifying potential occurrences and can be improved through information sharing about changes in advance of their occurrence. Anticipation is required to obtain knowledge about potential future changes (Zsidisin & Wagner, 2010). Anticipation feeds the proactive part of resilience by enhancing the ability to prepare for a disturbance, and therefore limiting its effect through robustness.

Preparedness, or ‘readiness’ directly contributes to resilience (Ponomarov & Holcomb, 2009), and encompasses all actions taken to prepare for the events identified in the ‘anticipation’ stage. Preparedness is based upon the ability to prepare an effective and efficient response to change (Altay & Green, 2006; Tomlin, 2006) and enhances resilience by increasing robustness through lowering the likelihood of a disturbance affecting the entire system.

By their very nature, preparedness and anticipation as described here are operational, or process based factors of resilience. A drawback of this conceptualisation lies in the fact that to effectively implement robustness, visibility (a factor of agility) is required, leading to the two being co-dependent. However, as this paper discusses the implementation of resilience, and thus robustness and agility simultaneously, this factor is deemed negligible in this setting.

2.1.2 Agility

(9)

Peck, 2004; Jüttner & Maklan, 2011; Pettit et al., 2010; Sheffi & Rice, 2005). Braunscheidel & Suresh, (2009) find that, flexibility and agility are conceptually distinct, though agility is a characteristic of flexibility (Christopher, 2000; Hohenstein, Feisel, Hartmann, & Giunipero, 2015). However, with agility defined as “the ability to respond rapidly to unpredictable changes” (Christopher & Peck, 2004: 18), and flexibility defined as “susceptible to modification or adaptation” (Dictionary.com, 2015), the researcher proposes that there is sufficient overlap in the definitions to enable the use of both words interchangeably, though for simplicity, agility will primarily be used in this paper, in line with the Wieland & Wallenburg (2013) conceptualisation of resilience used for this research.

Visibility, according to Christopher & Peck (2004), is the ability to ‘see’ up and down the supply chain. That is, a transparent view throughout the supply chain of inventories, supply and demand conditions, the status of other supply chain actors and so on. Visibility also implies clear lines of communication with other actors in the supply chain (Christopher & Peck, 2004, Barratt & Oke, 2007). Visibility in the supply chain is created through close collaboration with external supply chain parties and clear integration within the focal organisation. Christopher & Peck (2004) add that organisational discord is often a barrier to visibility in the supply chain as it obstructs the free flow of information, in turn reducing the visibility of risk (Harland, Brenchley, & Walker, 2003) resulting in reduced awareness of impending disruptions. This problem is compounded by lack of integration which Pal, Torstensson, & Mattila (2014) argue is also a contributor to resilience in organisations.

(10)

As incorporating supply chain resilience is a participative process (Scholten, Scott, & Fynes, 2014), and due to the intentional nature of resilience development, it is interesting to study resilience in combination with lean supply chain aspects, which are also intentional supply chain and organisational changes and processes. Also, as building resilience into a supply chain often comes at the (necessary) expense of adding redundancy or ‘slack’ into a supply chain (Sheffi & Rice, 2005), the contrast with lean provides an interesting trade-off in organisational decision making. Due to this, Christopher & Peck (2004) envision resilience as a trade-off between redundancy and efficiency.

2.2 Total Productive Maintenance

TPM is a proactive approach applied in anticipation of and to prepare systems, networks and organisations for disruptions. In an interesting parallel, Gits (1992), defines maintenance as any activity aiming to keep an item in, or returning it to the physical state considered necessary for the fulfilment of its function. Whilst this definition bears some similarity to that of resilience, it is not sufficient to equate maintenance to resilience, but rather to investigate how maintenance and resilience are linked. By exploring these linkages, this research aims to gain an understanding of how TPM can build resilience in a supply chain network.

(11)

organisational members to gain a deeper understanding of the ‘true cost’ of lack of resilience.

2.2.1 Antecedents of TPM

(12)

Table 2.1: Components of TPM in Literature

TPM Actions and Philosophies as found in literature

TPM Components Corresponding

Literature

1. Process and product quality are key to the performance of all employees.

2. Equipment failure can, and should, be prevented.

3. “If it ain’t broke, fix it anyway”.

4. Equipment performance can and should be managed.

 Ensuring participation in maintenance management:

o Point 1 indicates that making maintenance a priority across the organisation ensures participation in maintenance management.

 Increasing reliability:

o Point 2, 3 and 4 indicates that maintenance should be managed and carried out before it is required

(McKone & Weiss, 1998, (pp. 338))

1. Maximising equipment effectiveness.

2. Establishing maintenance strategies for the life of the equipment.

3. Include all departments in maintenance. 4. Involve all staff top-to-bottom.

5. Promote maintenance through small-group and autonomous activities.

o Point 1 emphasises the reliability of equipment through a measure of its contribution

 Preparing contingencies for disruptions:

o Point 2 encompasses the strategic nature of TPM as a proactive method of achieving resilience

o Points 3, 4 and 5 engage organisational members in TPM and emphasise its importance

(13)

1. Identification of autonomous and planned maintenance.

2. Equipment technology emphasis. 3. Committed leadership.

4. Strategic planning. 5. Cross-functional training. 6. Employee involvement

o Points 1 and 2 focus on enabling maintenance as and when is required

o Point 3 is important to show to all organisational members the level of commitment to TPM desired by the organisation

o Points 4, 5 and 6 allow for long term planning and to prepare for a wider range on disruptions through the incorporation of additional organisational members

( Nakajima, 1988;

Takahashi & Osada 1990; Tsuchiya, 1992;

Steinbacher &

Steinbacher, 1993; Maier, Milling, & Hasenpusch, 1998; McKone,

Schroeder, & Cua, 1999; Mckone & Weiss, 1998, in Cua, McKone, & Schroeder, 2001;), (Prabhuswamy, Nagesh, & Ravikumar, 2013) 1. Building profitable operations through reduction

of breakdowns

2. Practice ‘prevention is better than the cure’ 3. Involve all organisational members through

participatory management 4. Bring equipment to its ideal state

5. Create a self-sustaining cycle in the workplace based around culture and management.

o Point 1 makes a direct link between operational goals and longevity and TPM. Point 2 emphasises the importance of preventive rather than reactive maintenance

 Ensuring participation in maintenance management: o Points 3 and 5 show that ensuring participation in TPM is

both a top-down and a bottom-up process.

o Point 4 highlights that the best preparation for disruptions is to ensure equipment is always running ideally, and that it is easier to respond to disruptions with reliable equipment

(14)

Throughout literature on TPM, recurring themes are often attributed to Nakajima (1988), the so-called ‘father of TPM’ (Bamber, Sharp, & Hides, 1999; Cua et al., 2001; McKone & Weiss, 1998). Nakajima (1988) separates the purpose of TPM as improving the company through the improvement of its people and its equipment. This distinction is commonly addressed in TPM literature and it is therefore important to include it in any conceptualisation of TPM. The three key tenets of TPM found in the table above are explained next.

Increasing Reliability

TPM is an aid to improving maintenance in operations management (Prabhuswamy, Nagesh, & Ravikumar, 2013). By improving reliability, TPM reduces the occurrences of avoidable disturbances (Ahuja & Khamba, 2008). Reliability can be increased through prevention of equipment failure (Kleindorfer & Saad, 2005), and through the managing of equipment performance and technology (Prabhuswamy et al., 2013). Prabhuswamy et al. (2013) also state that reliability through maintenance helps to improve operations, as found by Nakajima (1988). These factors thus link increasing reliability to enhancing robustness, due to their contribution to reducing the possibilities of disturbances affecting a system.

Preparing Contingencies for Disruptions

(15)

Ensuring Participation in Maintenance Management (Continuous Improvement)

Arguably the most ‘familiar’ element of any lean philosophy is continuous improvement (Holweg, 2007). Categorised here as participation in maintenance management, involving organisational members in TPM initiatives is commonly found in literature on the subject. Additionally, it is important that application of TPM be all-encompassing in an organisation, including all departments, staff and through the creation of group activities to highlight to organisational members the importance of TPM (Nakajima, 1988) in the creation of a resilient network.

The three factors mentioned above, together indicate application of TPM in organisations.

2.3 Conceptual Model

The discussion of resilience above has shown that resilience is comprised of robustness and agility. Further, TPM actions and philosophies have been aggregated into three components. These three components highlight ways in which TPM contributes to increased reliability, prepares for disruption and ensures participation in maintenance management. Aligning TPM with the study of resilience suggests TPM can aid resilience by enhancing both robustness and agility, due to TPM containing both proactive and reactive elements.

(16)

Figure 2.1: Conceptual Model

3. Methodology:

In order to investigate potential relationships between TPM and resilience, and due to the specificity of the research question and its demand for detailed observations, a case study is deemed fitting (deMarrais & Lapan, 2003). Specifically a single case study is the appropriate method to glean in-depth and detailed information regarding the conceptual framework developed by the researcher to answer the questions stated in the introduction within the context of this research.

(17)

factors introduced in the previous section, this research aims to deliver an applicable component solution to the case organisation (Hevner, March, Park, & Ram, 2004). In order to do this, design science methods often seen in information system and engineering research (Peffers, Tuunanen, Rothenberger, & Chatterjee, 2007) are employed as part of the case study.

3.1 Introducing the Case

The focal company supplies water to private and business consumers throughout a province in the northern Netherlands and is the most suitable case context because of the importance of maintenance to achieving uninterrupted delivery of water to consumers (to provide a sense of proportion, the service level was found to be 99.9996%, with average downtime measured in minutes per annum). The organisation manages and maintains a network of water ‘factories’, pipelines and valves as part of its distribution network. To support the continued operation of this network the organisation maintains a network of storage locations, personnel, spare parts and tools. This support network is the singular focus of this study due to the importance of maintenance management to the organisation, and a desire to improve operational processes in this network whilst retaining a high service level (resilience), and the inherent challenges of maintaining a high service level whilst attempting to reduce redundancies. This service level is indicated by the amount of downtime (lack of water) experienced by customers as the result of a disruption. Thus in this setting, resilience is equated to minimisation of downtime caused by disturbances. A variety of restrictions hinder the ability of the organisation to engage in maintenance, namely the relative inaccessibility of the distribution network due to its location underground. It follows therefore that in the event of a disturbance, the processes required for a response should be both robust and agile.

(18)

provide a key example to which such a theory can be applied and investigated. Further, the organisation has an interest in understanding the effects of TPM on resilience, to gain insight into the importance of the design of their service network supporting their maintenance initiatives.

Despite gathering knowledge by medium of a single case study, the researcher would argue that this organisation is representative of an organisation dependent upon its supply network for maintenance and would suggest that this research is applicable to other, similar organisations in public (and semi-public) utilities, where there is a high necessity for management of asset-based networks and maintenance as well as a requirement for very high service levels against initiatives for cost reduction, such as in gas and electricity distributors, as well as other water companies.

3.1.1 Identifying the network

(19)

Figure 3.1: Simplified exemplification of current internal supply network layout

3.2 Problem Discussion

The context of the problem revolves around the desire of the organisation to redesign the supply network to achieve and maintain high resilience, in the form of minimised downtime. The substantial difficulty in this endeavour however, lies in the fact that a variety of stakeholders each holding different priorities and desired outcomes for this redesign have influence over the process. The problem can, in effect, be designated as a “wicked problem” defined as a class of problems that are poorly formulated, where the information is confusing, wherein many stakeholders hold conflicting views and values and where the ramifications of the system are complex (Buchanan, 1992; Rittel & Webber, 1973), or put simply “the formulation of a wicked problem is the problem” (Rittel & Webber, 1973: 161).

(20)

of discovered factors on service level and the links and effects they have on each other to aid the organisation in its decision making.

Design science methods are used due to their ability to improve the environment through the introduction of new and innovative artefacts (Simon, 1996), and their relevance to the case through an iterative process (Hevner, 2007) which refines the artefact during the research, based on environmental and theoretical influences as shown in figure 3.2, from Hevner et al. (2004), below. The iterative approach used in design methods research refers to the ‘assess and refine’ process visible in the centre of the figure, wherein the researcher develops models and theories for assessment, before accounting for further environmental and theoretical factors to increase the specificity of the theories and artefacts presented. The following section presents the methodology used to achieve this process in the case.

Figure 3.2: Design Science Research Framework (Hevner et al., 2004)

3.3 Methodology for solution development

3.3.1 Understanding the problem

(21)

gathering opportunity. Employees were selected with the help of a primary contact, based on their influence over the outcome of the decision, or because their role in the organisation would be directly affected by the decision. The archival data provided was used to gain additional background information regarding the organisation and the lead-up to this problem.

The researcher also spent several days per week at the focal organisation to facilitate access to organisational members and information, and informal conversations helped to clarify and improve the quality of the data gathered in formal interviews. Lastly, archival sources were examined (such as company reports).

Initial interviews were conducted with the aim to define the problem. To do so, themes discussed in an initial meeting were explored with interviewees to gain an understanding of the concerns and issues of stakeholders, and of the outcomes of the process desired by these stakeholders.

The problem defined was a lack of understanding of the effect on service level (resilience) of a potential redesign on the supply chain network supporting the maintenance of the organisation’s delivery network. Specifically, different organisational members with conflicting views and priorities advocated different potential layouts for the network and as a result, the organisation lacked an overall view of the ramifications of each potential layout, leading decision-makers to a stalemate, or in other words, the inability to objectively grasp and prioritise the key factors required.

3.3.2 Defining the objective

To aid the organisation in decision-making regarding the layout of its supply network, a comprehensive overview of factors affecting the decision was deemed necessary. Moreover, it became apparent that in isolation, decision factors presented little value, and it was therefore deemed necessary to present an understanding of how these decision factors were related, their interdependent effects and the overall effect the decision could have on the service level of the system. These decision factors would ultimately be presented as an artefact mapping the factors and their relationships. 3.3.3 Research and analysis activities

(22)

process involved in design science methods (Hevner, 2007), these two rounds served to refine the data gathered and ensure the applicability of the resulting artefact. In between the two rounds of interviews, a general meeting was held to inform participants of the progress of the project and to foster discussion around the issues discovered. Information gleaned in this meeting was used to provide direction for the second round of interviews. Table 3.1 provides a timeline and details of interviews.

Table 3.1: Details of meetings, dates and attendees

Meeting Meeting

identifier

Attendees Duration

Initial meeting Meeting 1 Primary contact, Head of Purchasing, Lean Specialist

35 minutes Refinement Meeting Meeting 2 Main contact, Research Supervisor 1 hour

Round 1 Meeting Meeting 3 Head of Purchasing 50 minutes

Round 1 Meeting Meeting 4 Head of Drinking Water 1 hour Round 1 Meeting Meeting 5 Head of Distribution Maintenance 35 minutes Round 1 Meeting Meeting 6 Head of Distribution for Projects 45 minutes Presentation and

discussion of findings to date

Meeting 7 Heads of Department, Research Supervisor, Primary Contact, Head of Water Supply

2 hours

Refinement Meeting Meeting 8 Head of Water Supply 55 minutes Round 2 Interviews Meeting 9 Head of Distribution for Projects, 50 minutes Round 2 Interviews Meeting 10 Head of Distribution Maintenance 40 minutes Round 2 Interviews Meeting 11 Head of Purchasing 1 hour Round 2 Interviews Meeting 12 Head of Industrial Water 35 minutes Round 2 Interviews Meeting 13 Head of Drinking water 1 hour Round 2 Interviews Meeting 14 Head of Water Supply 1 hour Round 2 Interviews Meeting 15 Asset manager Strategy and

Research

1 hour

Initial Phase

(23)

transcribed to Atlas.ti. Further investigation highlighted that ‘lean’ in its entirety was too broad for the scope of the research, whilst simultaneously it became clear that the company focused heavily on maintenance. Thus TPM was selected as a specific subset of lean for the investigation.

Problem and Objective refinement

The information gleaned in the initial phase helped shape the direction of the research and provided focus for the literature review and the finalisation of the theoretical model. A meeting with all interviewees and the primary contact was held to present the model and discuss the research direction. The aim of this meeting was to enable a discussion between all stakeholders and used as part of the data collection. Information from this meeting included recorded audio files and handwritten notes, both transcribed to Atlas.ti and referred to throughout the subsequent phases of the research to ensure the continued focus of the research activities. Lasting approximately two hours, the meeting was also used for approval of the model and to agree on the aim of the project (the deliverable artefact).

Data collection and case investigation

(24)

Quotations were initially classed by theme, based on the conceptual model, with the addition of a theme entitled ‘factors’, used to denote quotations which suggested potential factors of the decision-making process. Subsequently, the key ideas of the quotations were summarised in an adjacent column before adding two further levels of codes, each progressively being more specific at describing each quote. These codes were then grouped by topic based on these codes and analysed for (dis-) agreement, in order to balance arguments and provide a clear view of alternative responses (an example of coding can be seen in figure 3.3, and appendix C). When the meaning of an individual quotation was unclear, the researcher referred to the original transcript to evaluate the response within the context it was given in. Lastly, the quotations and corresponding codes were compared to the developed theory. The process was repeated and quotations and coding reviewed multiple times by the researcher to ensure data accuracy and fidelity.

Table 3.2: Details and use of archival data used

Information Use

Disturbance time reports To show disturbance times due to disruptions and downtime per customer

Report Pilot of "out of hours" workers

Used to show minimum number of workers required

Quarterly report Contains latest information on organisational variables including water quality, business performance and other KPIs

Long term planning report for network and distribution (two versions)

Contains information on the long term issues facing the business with regards to its infrastructure

The various supply chain network layout options (presented in appendix B) proposed by the company were used throughout the research to guide the discussion with interviewees regarding the potential contribution of these scenarios to TPM and, by extension, to resilience. The application of these layout options to the research was seen as vital by the researcher to bridge the gap between organisation theory and management theory as defined by Aken (2004).

Theme Quote Key idea First level code Second level code

Agility If there is a central warehouse only, there

will be excessive movement, which is a waste Increases overall distance travelled Speed Transport Requirement

(25)

4. Findings

4.1 Initial findings

The basic direction of the research was established in the initial meeting. To narrow the issues faced by the organisation, it was necessary to gain an appreciation of the meaning of ‘lean’ in the organisation, and to clarify the goals and requirements of the project. It became apparent that the majority of interviewees had come into contact with some form of lean initiatives, yet there lacked clarity in what these lean initiatives were aiming to achieve. This is evident from the variety of responses gained from asking interviewees to describe lean in their organisation, a selection of which is shown in table 4.1.

Table 4.1: Example of interpretations of lean across the organisation

Position Definition of Lean

Meeting 4 Lean is not a solution or tool, it is a mirror we use to ask ourselves what we are doing, how can it be done better?

Meeting 3 Lean is not a goal, it is used to improve work and do more with less. Meeting 5 Lean is a way to do things better, implementing steps to improve. Meeting 8 Lean is reduction of waste

When questioned, interviewees struggled to attribute their definition of lean to a concrete initiative. General consensus was that lean initiatives had been initiated top-down, but not sufficiently communicated to allow a bottom-up response from employees throughout the organisation, or that they were perceived as strategic concepts, for which there was insufficient time on a day to day basis, resulting in a lack of engagement and thus a lack of traction. In other words, the gains that such initiatives could yield for the organisation and its employees were not explicit, resulting in little buy-in.

Conversely, resilience was a topic largely unexplored by the organisation. However, upon explanation of the concepts behind resilience, consensus appeared that achieving supply chain resilience would be of interest to the parties involved in the supply network redesign.

(26)

maintaining a high level of service (a quantitative measure of resilience), whilst optimising activities in the remainder of the business.

4.2 Main findings

Upon hearing the definition of resilience put forward by the researcher, it was suggested that the service level of the supply network would equate to resilience in this context. “Focus on the service level of the maintenance and logistics systems, as

it would be difficult to translate how maintenance system improvements translate to improved customer minutes” (meeting 7). This quote points out an important factor

and a source of potential confusion, in that the key measure of service level in the organisation is ‘customer minutes lost’, whereas this study focuses on the service level of the supply network and maintenance processes, identified by “delivery minutes outside specification”. The difference is clearly highlighted in a quote from a later interview describing a situation where the customer minutes were not affected, but the supply and maintenance metric was: “In one example we had a leak in a

custom pipe, it needed replacing, but we could leave the leak until we were ready to fix the section, the customer noticed nothing” (meeting 11).

In this section, the findings regarding the effects of TPM on the antecedents of resilience are presented.

4.2.1 Robustness

As shown in the theoretical section, robustness is the proactive side of resilience, and is composed of anticipation and preparedness. In order to gather in depth data, these two entities were treated through separate questions in the interviews, so as to determine the effect of the redesign on both, and therefore, to determine the impact on robustness.

Anticipation

As anticipation revolves around the prediction of disturbances and eventualities, the effects of the redesign on anticipation revolved largely around knowledge creation, and information gathering in order to ‘predict’ possible occurrences and scenarios the supply network could be subjected to.

(27)

contribute to anticipation and preparedness simultaneously. For example, the quote

“we have a matrix based on urgency and importance, and based on the position of various materials in the quadrants, we know at what point in the supply chain we can position certain items” (meeting 13) covers both as the determining of urgency and

importance lies squarely under anticipation, but acting on that knowledge to place items at various points further up- or downstream falls under preparedness. Nevertheless, the tenet of TPM contributing most strongly to anticipation was found to be the preparation of contingencies for disruptions. A prime example of this is the tendency of the organisation to order over and above projected material requirements when engaging in project work (planned repairs and maintenance). “Now we put 15%

extra into the plan, […] but sometimes we need something else that we had not planned for, so maybe 20% of calculated material is going back” (meeting 9). This

kind of contingency planning is considered normal practice based on knowledge and experience gained in previous project work. This is caused by the deviation of actual requirements from planned requirements based on information available in plans and observations made after work begins (for example, when pipes below ground do not match drawings). To this end, activities with an eye to improve reliability are aimed at increased anticipation in turn leading to increased preparedness. In this case, efforts are being made to reduce and standardise the number of components and parts used in maintenance and repair work (by using parts falling under such standards as DIN and EN1). This is done in anticipation of future maintenance work as part of a long-term plan, and as a way to simplify tasks and procedures, enabling more standardised procedures and processes, both improving the anticipation of future requirements, and reducing the need for anticipation and shifting towards increased preparedness. Utilising standardised parts has the added benefit of enabling the organisation to procure materials from a wider selection of suppliers, potentially increasing the availability and speed of delivery of material.

Employee involvement also contributes to anticipation. In the organisation, emphasis is placed upon utilising historical data and knowledge to anticipate future occurrences. Disruptions in the distribution network are often based on degradation of pipes, pumps and valves, due to age, location and environmental factors. Whilst the

1

(28)

data can be used to measure reliability and prepare contingencies, it only gains value through interpretation by employees and specifically by the operators who have in-depth knowledge and expertise in dealing with these materials. They are best placed to participate and provide input, which helps to build contingencies and anticipation. This combined information is put to use to estimate the likely lifespan of the rest of the network, enabling the organisation to anticipate failures, “one issue we cannot

forget, is that the longer an item has been in service, the more likely it is to fail, but we cannot go and check” (meeting 15). Where such information is not available, or

predictions cannot be made with certainty, as is the case for “pipes that were laid in

the sixties, with an expected lifespan of 100 years” (meeting 15), the organisation has

this year, enacted a 50-year plan. This plan anticipates future requirements and is based on a concerted effort to gather data and expertise. Further use of employee involvement is made in the building of working standards and operating procedures both in practice and through initiatives aiming to classify and ‘inventorise’ the knowledge of workers. This initiative aims to define the skill level and training of each worker for recourse when future disturbances call for specific expertise, enabling a more efficient response to disturbances. This is exemplified by the following quote,

“I want people to spend time on development, on think work, not on digging holes, but on coming up with ideas to improve the processes, to make things simpler, and when things get simpler, the chance of faults comes down” (meeting 14); this

sentiment is echoed by another interviewee aiming for greater versatility in the workforce, whilst overcoming the division of roles perceived by older workers, “not

to ask yourself whether it is in your functional profile. If you can do it, get it done. But the problem is the culture, and older workers joined the company many years ago, to do one job, and unlike younger guys, they don’t want to take ownership of a problem which isn’t ‘theirs’” (meeting 13).

Whilst anticipation may not directly have an impact on the redesign of the network, the organisation is aware that the knowledge created in the anticipation phase, has an input on the redesign as it allows the organisation to gain a partial understanding of the demands which will be placed on the network in the future, and balancing these with immediate demands, as shown in: “we need to be realistic and ask ourselves

when, but also how often will a disturbance occur? And plan accordingly” (Head of

(29)

Preparedness

As the more practical side of robustness, preparedness is more readily ‘visible’ as an influence on the redesign, and usually stems from factors identified in the anticipation stage. Preparedness is evident in the supply network through the placement of specific parts and tools nearest to areas most likely to require them (in the current system), as well as periodical checking of materials to ensure they are fit-for-use. Such activities are reinforced by TPM factors. By periodically inspecting materials, valves and pumps, the reliability of both the distribution and supply networks increases, prompting an increase in preparedness. The placing of tools and parts at strategic points throughout the network prepares the network for contingencies. Effectively, in this network preparedness as contributing to robustness and preparing contingencies for disruptions as defined for TPM are equal, and both are strongly influenced by information and knowledge gathered and acted upon in the anticipation phase. The outcome of this being that in order to enhance both anticipation and preparedness (thus, robustness), visibility is also required. Simply put: “when you have good basic

information, and you know what is in stock, you improve robustness” (meeting 13).

Combined with better information management, one aspect of the long term view of the network entails that one-off and custom parts be phased out long-term, reducing the latency currently in the system whilst required parts and expertise are gathered to perform maintenance and repairs. In time, this will reduce rather than efface the need for contingency planning, yet enable the organisation to be more specific and effective in planning more robust processes for a smaller array of contingencies. Robustness is included in the redesign with the aim to affect the readiness of processes in the face of future disturbances. This is exemplified by the generation of knowledge regarding maintenance activities, which can be recorded and referred to as a standard operating procedure (SOP), improved over time, leading to more dependable processes through the involvement of employees and the preparation for contingencies: “they [the workers] can record and instruct how repairs should be

done, and describe how we can do things better next time” (meeting 14). Coupling

(30)

contingencies for disruptions and ensuring employee involvement affects anticipation and preparedness, TPM has made apparent the need for information about the supply network to improve resilience.

4.2.2 Agility

When explained during the interviews, interviewees acknowledged that agility was a key attribute to consider when developing the renewed supply network, though further explanation was required in some cases to highlight how visibility and velocity contributed to agility. Following such an explanation, the two factors were treated through separate questions in order to delve deeper into the topic and gain an impression of the effect of TPM and the redesign on agility.

Visibility

In this context, visibility transpired to be the key antecedent of resilience, and one that would enable improvements in all other factors. Visibility was attributed to two main factors, both concerned with information and knowledge. The first considered the physical properties of the distribution network and the availability of knowledge regarding it; that is the accuracy of the information in the system compared to what parts and components physically lay beneath the ground. The second factor regards information pertaining to stocks, inventory and the movements of parts, tools and people within the network.

The organisation is enacting processes to increase the reliability of both the network and the processes revolving around the network and its maintenance. Due to the historical nature of a large portion of the distribution network, inaccuracies in information due to out-dated drawings, and gaps in information regarding the physical layout of water mains (e.g. the position, depth, material, etc.) abound. Increasing the reliability of the information available to workers is an on-going process, facilitated by recent investments in technologies such as the deployment of smartphones, tablets and dedicated applications through which workers at work sites can add information, images and update technical drawings as they encounter issues and execute repairs.

“The drawings we have of what is on-site, are not always accurate, so you can arrive with the wrong materials and tools. This improves with the iPad app” (meeting 14).

Visibility of the distribution network will improve over time as the network is worked on and the knowledge of the system updated: “Now, with technology, you can notify

(31)

accuracy over time” (meeting 10). This visibility enables more effective planning of

part requirements, allowing the organisation to make more informed layout decisions. A key issue multiple interviewees identified, and aim to review through the redesign was a lack of visibility in the inventory and stock management systems the organisation uses. To highlight the extent of the issue, a selection of quotations from interviews is shown in table 4.2.

Table 4.2: Quotations on stock and inventory control in the system

Quote Key idea Second

level code

First level code There are several paths from suppliers to the

worksite, direct, through a work order, through a central location and out and so on. The problem is, if one happens, the other systems are not informed, leading to wasted time and duplicated work.

Opening multiple channels without a unified view leads to a loss of control

Systems Problem

Visibility

We have to create an internal system where the central warehouse and the rest of the system communicate Increase accuracy and communication Systems solution Visibility

Everyone should be able to see what is in stock Information accuracy improves communication Systems solution Visibility

We must get the basic information right Knowledge improves planning of work

Legacy problem

Visibility

Because of the satellite layout, we do not know exactly where things are

Lack of control over processes increases misinformation Legacy problem Visibility

Currently, there is no data exchange between the central warehouse and the sub-warehouses. So once something leaves the central

warehouse it disappears. Lack of control over processes increases misinformation Legacy problem Visibility

Due to lack of visibility, we cannot see parts, so we assume they are not there so we order more Lack of control over processes increases misinformation Systems Problem Visibility

A better future system is that "satellite locations" (factories and pumping stations) have only tools and machinery, no

consumables, so we have no stock visibility problems

Reduce errors in information by removing the cause of misinformation

Layout constraint

Visibility

There is no connection between the maintenance system and the logistics and inventory system. So it's difficult to assign material to work. Traceability of work and requirements Systems Problem Visibility

There are a lot of parts we know they are there but we don't know if we can take them or if they are reserved

(32)

Increasing reliability, and specifically increasing the reliability of information was discussed as a predominant force driving the perceived need for the redesign. Lack of reliability is emphasised by gaps in information regarding the management of inventory. This is due to limitations in the information systems that govern the current central warehouse. The tracking of parts that leave the central warehouse (for example, for storage in vans, or at sub-warehouses) is not possible, leading to large amounts of ‘lost’ inventory. This causes knock-on effects and irregularities in the procurement systems as this limitation does not allow for tracking of actual usage, but merely shows availability in stores, as explained: “for example if the worker takes a

box of 100 screws on day one, they leave the system, and the system will order 100 more screws on day two, but the worker on average uses one screw a day, meaning that the new box will sit on the shelf for 99 days, and was ordered maybe 90 days too early, if we could track actual use, this is where we can make small savings that all add up” (meeting 13). Despite appearing to be a systems problem, it was voiced by

some interviewees that these systems can be solved by simpler methods, for example through (as advocated by a TPM approach) empowering employees to self-manage the reordering of small parts that are not centrally stored or through the implementation of non-IT systems, “Another option would be to use some kind of

Kanban2 system for sub-locations. It's simple, cheap and efficient and doesn’t require complex processes” (meeting 13).

The fact that there is widespread awareness of the issue engenders consent that the issue should be addressed in the redesign, or conversely that the redesign provides the organisation with an opportunity to also redesign systems to manage and control inventory, depending on the ultimate design selection.

Velocity

The concept of velocity is crucial to the redesign, particularly due to the main metrics used to quantify the service level of the distribution network, customer minutes lost, and the metric used to determine the service level of the supply network, out of specification delivery minutes, and the impact of the speed of response on that metric. However, answers to questions about speed mostly focused on the driving time and distances of operators that would result from potential network layout changes.

2_{Kanban is a lean logistics and inventory system comprising utilizing boxes and cards to pull material}

(33)

“There is an upper limit to speed, due to the act of driving to the disturbance, and more people will not improve that” (meeting 10). Further, layout changes are not

expected to change the speed of work preparation, much as they will not have particular impact on supply chain processes linked to the purchasing department, as these processes are more dependent on other system limitations such as processes, stock management and operational procedures, but instead that the layout will be a key determinant of speed “Layout has no effect on work preparation, but for speed it

could do” (meeting 11).

Despite such affirmations, other stakeholders are of the opinion that speed could be improved due to changes in layout stemming from procedural changes such as the keeping of stock in vans, or the creation of a separate logistics team, with the sole role of supporting workers in the field in delivering tools and parts: “Consumables should

be in the vans and the central warehouse, and registered on the iPad when they are used, that way the vans are replenished at work sites by a logistics only van, that way people can repair things and not drive all day” (meeting 11), effectively removing

the need for workers to return to a central (or sub) location in between tasks, reducing time and distance, in effect increasing the speed of response.

Overall, speed is most affected by the layout selected; however speed was perceived as less of a concern to interviewees than, for example, the cost associated with driving time and distance that would result from the different layout options which interviewees agreed should aim to be minimised. At current service levels, interviewees were of the opinion that speed was sufficient and gains in other areas would be more beneficial to the resilience of the network. The three tenets of TPM were also found to have little contribution to speed, though forward planning by preparing for contingencies would aid speed of response due to the potential for reduced distance driven on account of the higher likelihood of workers having required materials available, rather than needing to collect them from stock.

4.2.3 Resilience

Building on decisions made during the meeting prior to the second phase of the research, wherein the service level of the logistics network would relate to resilience, questions revolved around how the redesign would affect this metric and why. One interesting view, differing from that of colleagues was voiced by an interviewee, “the

(34)

let’s call them costs, of other factors” (meeting 15). Whilst the interviewee was

referring to customer service level, rather than supply network service level, this quotation is emblematic of the aim and primary preoccupation of the organisation. Questioning and discussing the effect of TPM factors on the network and redesign led to the conclusion by interviewees that the factors would lead to increased resilience in the form of: on time materials at work sites, reduced variability in lead times in the supply chain and the network, decreased response time for repairs from initial awareness, and reduction of stock-outs and duplicate orders within then network. The availability of such metrics will enable the organisation to identify and measure the network benefits of the redesign. These can be aligned with fundamental metrics mentioned in a later interview. The water quality index and an out of specification delivery minutes measure: “the OLMs and WKIs are, I think, the main things

affecting the way we work” 3

(meeting 14).

Identifying how each proposed scenario would impact resilience based solely on qualitative data remains challenging. Equally, finding a balance between ‘hard’ data such as the metrics mentioned above and ‘soft’ data such as the factors mentioned in the theoretical model adds to the challenge of measuring the overall resilience of the network. At a basic level, ‘hard’ data also includes such measurements as costs incurred by the redesign and costs incurred as a result of the redesign, such as investments in technology, and infrastructure costs (building or adapting of new and existing facilities), as well as costs incurred due to potential layout changes, such as stock and inventory costs as well as costs incurred due to distance and time spent driving by operators. Despite this, in this context, engaging in the process of improving resilience through TPM has raised important questions regarding the network for the organisation, and highlighted key areas for improvement.

5. Discussion

Despite the origins of TPM being firmly rooted in manufacturing (Ahuja & Khamba, 2008), revolving around practices geared to reduce the likelihood of breakdowns and improve the longevity of machinery and production lines; TPM has been proposed here as more than an improvement methodology comprised of specific steps as

3

(35)

defined by Bamber, Sharp, & Hides, (1999) and Pramod, Devadasan, Muthu, Jagathyraj, & Dhakshina (2006). In this research, TPM is used as a contributor to supply network resilience by way of the antecedents of resilience proposed by Wieland & Wallenburg (2013). The proposed framework was evaluated in the context of a case and used to create a model enabling the organisation to determine the resilience of different supply network layouts considered by the focal company as part of a redesign of the supply network. The underlying mechanism, explaining how TPM can enable resilience by way of robustness and agility is elaborated and linked to theory in the following sections.

5.1 TPM as an enabler of robustness

(36)

declines in the face of complexity (Comfort, Sungu, Johnson, & Dunn, 2001). On an inventory management level, the organisation is already making a conscious effort to reduce the number and variety of parts in use in the network, proactively reducing complexity, however such drives should be translated to systems levels to ensure consistency throughout the network and to avoid problems such as those found in current replenishment practices, as described previously, and as advocated by (Craighead, Blackhurst, Rungtusanatham, & Handfield, 2007).

By gathering and interpreting data about the system, anticipation aids in determining the potential layout of the network due to influencing the potential locations of stocks and tools, preparing the network for future disruptions. In practical terms, and contrary to the model formulated to help guide the research, visibility was paramount to enable robustness. This emphasises the shortcoming of the conceptual model of resilience by Wieland & Wallenburg (2013), used in this paper to formulate the theoretical model, wherein the definitions of preparedness and anticipation are highly operational, and the definitions for velocity and visibility much more concrete, though confirms the positive influence of relational competencies, such as cooperation and communication between supply chain partners (or in this case, between actors in the supply network) as the findings by Wieland & Wallenburg (2013) confirm, as noted by Hohenstein et al. (2015). As anticipated, visibility is required to effectively implement robustness, making it challenging to ascertain the singular effect of robustness on resilience without also considering agility.

5.2 TPM as an enabler of agility

At the outset of the research, it was expected that TPM would have a more discernable effect on robustness, and that benefits arising from the improvement of robustness would have an overall positive, if not marked, effect on agility. This was partially due to the expectation that the results would concern operational and process level changes rather than information management (visibility). This initial misconception emphasises the challenge inherent with formulating wicked problems, as defined by Buchanan (1992) and Rittel & Webber (1973).

(37)

network, is a realisation of the extent to which lack of visibility and clarity hampers the resilience of the network. Exploring visibility in the supply network has led to uncover gaps in the knowledge of the organisation regarding its delivery systems, impeding their ability to adequately plan and prepare for disruptions, confirming suggestions by Mckone & Weiss (1998) wherein performance of a system can and should be managed. However, it is through attempting to improve the system that these shortcomings are identified, leading to improved understanding of the systems available and their limitations, as can be achieved through TPM according to Chan et al. (2005).

(38)

5.3 Resilience as a context-specific artefact

In the context of the case and the supply network redesign, resilience is comprised of robustness and agility, as found by Wieland & Wallenburg (2013). However, building an artefact to aid decision-making necessitated the addition of elements of TPM gathered from Ahuja & Khamba (2008), Bamber et al. (1999), Chan et al. (2005), McKone et al. (1999), Nakajima (1988) and Suzuki (1992), amongst others, as tabulated in section two and formed into tenets. The basic framework used to craft the research instrument allowed the gathering of additional important considerations for the decision making process such as costs, a key factor in the decision for the organisational members interviewed. The ensuing artefact is colour coded to denote factors of resilience in blue, factors of TPM in red, and to contrast between ‘hard’ and ‘soft’ metrics, identified by hashed and solid fill (colour) respectively. Shown in figure 5.1, the artefact is inspired by universal modelling language notation, each category contains key contributions made by that category to others. Arrows denote the linkages, with positive and negative symbols explaining the relationships. As an example, layout was found to have a positive or negative effect on speed, denoted by an arrow carrying a ‘-/+’, and visibility was found to increase anticipation, denoted by an arrow bearing a ‘+’.

This diagram simplifies the decision model by aggregating the decision-making factors required (a complete model showing every link identified is shown in appendix D). Overall it can be seen that visibility becomes a prerequisite factor in utilising TPM to enable resilience in a supply network. The model highlights the need to understand the systems surrounding the supply network in their entirety to make an informed decision regarding network layout. The complexity found throughout the formulation of the (wicked) problem explains the differences in the opinions of stakeholders regarding priorities for the redesign.

5.5 TPM as an improvement tool

(39)

(40)

(41)

6. Conclusion

This research highlights interesting results, contributing to theory and new insights into the effect of TPM on resilience in a supply network context. Whilst the initial research ideas were formed around much more practical and operational concepts, the results of this research are much more nuanced than was originally expected. The concepts applied blend highly theoretical constructs such as resilience, with concepts intended for practical application. This mixture applied itself particularly well to the study of a physical system, largely controlled by intangible processes and procedures. Moreover, the process applied to this research, inspired by research methodologies from information and design science systems research leant itself particularly well, as the resulting findings play heavily on the management of information and combine the management of information with the management of components and tools, showing the requirement for conformity in both a physical and a virtual setting. To discern and answer how applying the use of TPM could build supply network resilience, this study deconstructed TPM into three key tenets and investigated their interaction with antecedents of resilience, thus highlighting their effect on resilience. These interactions were investigated in a supply network setting and TPM was shown to have the potential to enable resilience, with the caveat that visibility is the driving factor in enabling resilience in a supply network. This research practically contributes to the case organisation, and similar organisations operating in similar contexts by developing a decision-making artefact, clearly highlighting the key factors to consider when redesigning a network layout with the aim to achieve resilience, as well as by suggesting the use of the tenets of TPM defined in this paper, to investigate internally where gains in resilience can be made and how.

6.1 Managerial implications

(42)

prepared for disturbances is key to providing customers with high service levels. Anticipating and preparing for disruptions hinges on managing knowledge and visibility of information in the organisation and the network. Managers attempting to enhance resilience using TPM will quickly find that information is required to do so, and it is therefore advised they start the process by identifying available information regarding their network and gather an understanding of additional information required. It is also important to remember that actors in the supply network (employees), are often best placed to provide input regarding these processes and that therefore, managers should involve employees from across the organisation in resilience initiatives.

6.2 Limitations and suggestions for future research

(43)

References:

Ahuja, I. P. S., & Khamba, J. S. 2008. Total productive maintenance: literature review and directions.

International Journal of Quality & Reliability Management, 25(7): 709–756.

Aken, J. E. Van. 2004. Management Research Based on the Paradigm of the Design Sciences : The Quest for Field-Tested and Grounded Technological Rules. Journal of Management Studies, 41(2): 219–246.

Altay, N., & Green, W. G. 2006. OR/MS research in disaster operations management. European

Journal of Operational Research, 175(1): 475–493.

Bamber, C. J., Sharp, J. M., & Hides, M. T. 1999. Factors affecting successful implementation of total productive maintenance: A UK manufacturing casre study perspective. Journal of Quality in

Maintenance Engineering, 5(3): 162–181.

Barratt, M., & Oke, a. 2007. Antecedents of supply chain visibility in retail supply chains: A resource-based theory perspective. Journal of Operations Management, 25(6): 1217–1233.

Blackhurst, J., Dunn, K. S., & Craighead, C. W. 2011. An Empirically Derived Framework of Global

Supply Resiliency, 32(4): 374–391.

Brandon-Jones, E., Squire, B., Autry, C. W., & Petersen, K. J. 2014. A Contingent Resource-Based Perspective of Supply Chain Resilience and Robustness. Journal of Supply Chain Management, 50(July): 55–73.

Braunscheidel, M. J., & Suresh, N. C. 2009. The organizational antecedents of a firm’s supply chain agility for risk mitigation and response. Journal of Operations Management, 27(2): 119–140. Buchanan, R. 1992. Wicked problems in design thinking. Design Issues, 8(2): 5–21.

Burnard, K., & Bhamra, R. 2011. Organisational resilience: development of a conceptual framework for organisational responses. International Journal of Production Research, 49(18): 5581– 5599.

Chan, F. T. S., Lau, H. C. W., Ip, R. W. L., Chan, H. K., & Kong, S. 2005. Implementation of total productive maintenance: A case study. International Journal of Production Economics, 95(1): 71–94.

Christopher, M. 2000. The agile supply chain: competing in volatile markets. Industrial Marketing

How can Total Productive Maintenance Enable Supply Network Resilience? A Case of Supply Networks in Utilities

Master Thesis Supply Chain Management

University of Groningen